[tin-dev] better kill-file feature?

Urs Janßen urs at tin.org
Wed Jan 20 10:01:06 CET 2016


In <mailman.434.1453275396.23414.tin-bugs at tin.org>, Ulli Horlacher wrote:
> In my favourite news groups there is a lot of off-topic nazi spam, like
> this one:
> 
> Path: news.in-ulm.de!newsfeed.in-ulm.de!news1.tnib.de!feed.news.tnib.de!news.tnib.de!newsfeed.freenet.ag!87.79.20.101.MISMATCH!newsreader4.netcologne.de!news.netcologne.de!goblin3!goblin.stu.neva.ru!news.netfront.net!regionalpo
> From: regionalpo at gmx.at (Yvonne Regionalpo)
> Newsgroups: de.soc.politik.misc,de.talk.tagesgeschehen,de.rec.fahrrad
> Subject: Regensburg =?ISO-8859-1?Q?ber=E4t_=FCber?= "Lady-Zonen" in Bussen
> Date: Wed, 20 Jan 2016 02:10:03 +0200
> Organization: GMX
> Lines: 35
> Message-ID: <1mhawmf.eud8j8qc0hglN%regionalpo at gmx.at>
> 
> 
> They all come from the open news server news.netfront.net (russia?).
> I cannot put it in the killfile, because From, Subject and Message-ID
> change with each post. 
> 
> My request now is: I want to select/kill based on Path or Newsgroups field
> (I am not interested in de.talk.* at all).

tin can just filter on (most of the) the std-headers available in the
NOV files (Subject, From, Date, Message-ID, References, Bytes, Lines and Xref)

as everthing else would result in
a) (much) more traffic
b) noticeable slow down of the filter

depnding on the server one might end up in doing a HEAD for each article in
the group. that would be number_of_articles commands/responses instead of
one and each response is usualy 2-5 times bigger than the related
data in the [X]OVER response. (LIST HEADERS/HDR (or XHDR) would help
if the server suppprts if for that header.)

it is on the TODO list for 15-20 years now

| o  fix kill/hot mechanism so that it is possible to match articles on
|    arbitrary headers

so I doubt it will be implemented within this decade. 

@tin-dev: any volunteers ,-)

urs
-- 
“That was a great time, the summer of '71 - I can't remember it, but
 I'll never forget it!” Lemmy Kilmister



More information about the tin-dev mailing list